Identification of moving vehicle parameters using bridge responses and estimated bridge pavement roughness. (15th December 2017)

• Record Type: Journal Article
• Authors: Wang, Haoqi; Nagayama, Tomonori; Zhao, Boyu; Su, Di
• Is Part Of: Engineering structures. Volume 153(2017:Dec. 15)
• Page Start: 57
• Page End: 70
• Publication Date: 2017-12-15
• Subjects: Vehicle-bridge interaction -- Particle filter -- Pavement roughness estimation -- Weigh-in-motion; Structural engineering -- Periodicals; Structural analysis (Engineering) -- Periodicals; Construction, Technique de la -- Périodiques; Génie parasismique -- Périodiques; Pression du vent -- Périodiques; Earthquake engineering; Structural engineering; Wind-pressure; Periodicals; 624.105
• DOI: 10.1016/j.engstruct.2017.10.006
• Languages: English
• ISSNs: 0141-0296

Abstract

Highlights: A particle filter approach estimates bridge pavement roughness using vehicle responses. Vehicle parameters are identified from bridge responses and the estimated pavement roughness. Bridge displacement estimates are included in the observation to improve the accuracy. The weighted global iteration method improves the parameter identification. Abstract: Passing vehicles cause bridge deformation and vibration. Overloaded vehicles can result in fatigue damage to, or even failure of, the bridge. The bridge response is related to the properties of the passing vehicles, particularly the vehicle weight. Therefore, a bridge weigh-in-motion system for estimating vehicle parameters is important for evaluating the bridge condition under repeated load. However, traditional weigh-in-motion methods, which involve the installation of strain gauges on bridge members and calibration with known weight truck, are often costly and time-consuming. In this paper, a method for the identification of moving vehicle parameters using bridge acceleration responses is investigated. A time-domain method based on the Bayesian theory application of a particle filter is adopted. The bridge pavement roughness is estimated in advance using vehicle responses from a sensor-equipped car with consideration of vehicle-bridge interaction, and it is utilized in the parameter estimation. The method does not require the calibration. Numerical simulations demonstrate that the vehicle parameters, including the vehicle weight, are estimated with high accuracy and robustness against observation noise and modeling error. Finally, this method is validated through field measurement. The resulting estimate of vehicle mass agrees with the measured value, demonstrating the practicality of the proposed method.